Polls didn’t get it wrong; the Tory campaign got it right.
Sunday, May 10, 2015 at 04:06PM
WTM

by Peggy Papakosta, @PeggyPapakosta

The polls or rather the data was not wrong in the run-up to the UK election on May 7th 2015. Polling analysis and their data can be nothing more than a snapshot of the moment. That’s not the problem if (1) all (2) raw data is (3) honest and (4) public. The problem begins when pollsters try to make the necessary adjustment to predict the final result. In order to do that, they split the ‘undefined’ vote (which is made of the undecided vote and the refusers) mainly between the two major parties.

In the Scottish Referendum, the yes or ‘leave the UK’ vote got 87% of the undecided. An amazing feat for the Scottish National Party.  If all the raw data was honest and published during the last week before the election, then the conservatives managed to get about 67% of the ‘undefined’ vote and the labour party only got about 24% (table below). An equally amazing feat for the tories. No one expected it. Not even the tory leadership. Usually pollsters follow the unwritten rule dictated by statistical experience and split the ‘undefined’ vote between the two big parties correspondingly to their unadjusted voting intention percentages. In this UK election, the unwritten rule was proven wrong because there was a big swing towards the conservative party at the last minute and the polls did not have time to pick it up. It is entirely possible that a campaign is so successful that it defies the unwritten rule of splitting the undecided and refusers somewhat evenly. It is probable that that the Tory campaign may have done a hell of a job and got most of the undecided vote. Some say that the potential informal or formal coalition of the Labour party with the SNP pushed the English reluctantly to the conservative embrace. And that could only be a success of the Tory campaigners that promoted this message relentlessly.

A simple way to go forward is for all pollsters to publish all their raw data without adjustments that exclude the undecided and the refusers and include both phone and internet polling. And let others decide where the swing will go and by how much.

Poll

Pollster

Conservatives Voting Intention 2015

Labour

Voting Intention 2015

Don't know

Refused

Don't know+ Refused

5&7 May Populus

 

27

29

9

2

 

5&6 May

Ashcroft

23

24

9

9

 

5&6 May

ComRes

27

27

12

6

 

05-May

BMG

28.2

30

11.3

2

 

05-May

Survation

28.6

30.5

9.2

1.1

 

3&6 May

ICM

22

25

15

10

 

3&5 May

ComRes

27

25

15

5

 

1&3 May

Ashcroft

21

21

11

10

 

1&3 May

Populus

27

27

9

2

 

02-May

Survation

26.9

30.1

12

0.7

 

01-May

Survation

27.7

29.9

12.4

1.6

 

28&30 Apr

ComRes

24

24

21

4

 

29&30 Apr

Populus

26

27

11

2

 

24&26 Apr

Ashcroft

25

21

9

10

 

Average of polls during last week

 

25.7

26.5

11.9

4.7

16.5

Election Results 2015

 

36.9

30.4

 

 

 

Results minus Average

 

11.2

3.9

 

 

 

% of Don't know + Refused

 

67.5

23.8

 

 

91.4

 

Article originally appeared on (http://whitetowermedia.com/).
See website for complete article licensing information.